What’s new in Grafana Cloud

You can now configure the Azure Monitor data source to authenticate as the logged-in Grafana user when making query and resource requests if you also use Azure Entra to sign your users into Grafana.

Current User authentication allows you to enforce Azure RBAC restrictions on your Grafana users by removing the need to provide broad service credentials. Once a data source is configured with Current User authentication a user will only have access to resources they can access directly in Azure.

We’ve made a number of improvements to the canvas visualization.

Enhanced flowcharting functionality

With the newly added Infinite panning editor option, you can now view and navigate very large canvases. This option is displayed when the Pan and zoom switch is enabled.

To try out this feature, you must first enable the canvasPanelPanZoom feature toggle.

The new alert rule detail view has a new look and feel with helpful metadata at the top. The namespace and group are shown in the breadcrumb navigation. This is interactive and can be used to filter rules by namespace or group. The rest of the alert detail content is split up into tabs:

Query and conditions

We released generative AI support in both OSS and Cloud last year, supporting features like Incident auto-summary and dashboard creation assistance. One downside was that it was tedious to set up – you had to install & enable the LLM plugin, and separately sign up with OpenAI or Azure to configure your own API keys.

On Cloud, this just got a lot easier!

The Grafana plugins catalog contains over 200 data sources, panels and applications to help you make the most of your data. However, finding the right plugin can sometimes be a challenge.

Our initial search capability required an exact match on the name of the given plugin.

In Grafana v10.1, we added a Tempo search editor powered by TraceQL (search tab). We also recommended using this new editor over the older non-TraceQL powered editor.

The older non-TraceQL powered editor has been removed. Any existing queries using the older editor will be automatically migrated to the new TraceQL-powered editor.

(Re-)introducing “Keep Last State” to Grafana managed alert rules.

You can now choose to keep the last evaluated state of an alert rule when that rule produces “No Data” or “Error” results. Simply choose the “Keep Last State” option for no data or error handling when editing a rule. Refer to the Alerting documentation on state and health of alert rules for more information.

We are happy to announce Cloud Logs Export is now Generally Available to all Grafana Cloud users!

Cloud Logs Export lets users ship logs to their own object storage (AWS, GCP, Azure) for low-cost, long term retention. If you need to retain logs for long periods of time, without a need to frequently query them, Cloud Logs is the most cost effective way to do that.

Dashboards, when accessed by users with the Viewer role, are now using the Scenes library. Those users shouldn’t see any difference in the dashboards apart from two small changes to the user interface (UI): the variables UI has slightly changed and the time picker is now part of the dashboard container.

Dashboards aren’t affected for users in other roles.

Directly from Kubernetes Monitoring, you can click into the Asserts RCA Workbench to see what is occurring between components and their relationships.

(Release 2.0.4)

You can now use generative AI to assist you in your Grafana dashboards. So far generative AI can help you generate panel and dashboard titles and descriptions - You can now generate a title and description for your panel or dashboard based on the data you’ve added to it. This is useful when you want to quickly visualize your data and don’t want to spend time coming up with a title or description.

Make sure to enable and configure Grafana’s LLM app plugin. For more information, refer to the Grafana LLM app plugin documentation.

Flamegraph AI in Grafana Cloud Profiles is now GA

Flamegraph AI helps you understand your profiling data by using a large-language model (LLM) to assist with data interpretation.

On any Pod or workload detail page, you can navigate directly to Application Observability to see more data about the application. (Release 2.0.3)

When you use the easy configuration process with Grafana Kubernetes Monitoring Helm chart, you can switch on the use of OTLP and Zipkin receivers for containerized application monitoring.

(Release 2.0.3)